Commit graph

726 commits

Author SHA1 Message Date
Edward Hervey
c138aa8648 collectpads: Make sure the CollectData list is up-to-date when reading/setting it
Without this, we risked:
* Checking the flushing state on an unexisting list
* Not setting the flushing state on pads that had just been added

Partially fixes #590056
2009-07-29 11:50:06 +02:00
Edward Hervey
fcbba9b15a collectpads: Split out _check_pads into a version without lock taking.
This is so we can use _check_pads in places where we've already taken
the lock in question.

Partially fixes #590056
2009-07-29 11:49:39 +02:00
Robin Stocker
527da05476 basesrc: don't handle SEEKING queries for formats that don't match the one the source operates in
Return FALSE in basesrc's default query handler when we get a SEEKING query for
a format that's not the one the source operates in. Previously (ie. before, in
the git version) we would return TRUE in that case and seekable=FALSE, which
is more correct, but causes backwards compatibility problems. (Before that
we would change the format of the query when answering, which was completely
broken since callers don't expect that or check for it). Since the SEEKING
query is a fairly recent addition, not all demuxers, parsers and decoders
implement it yet, in which case any SEEKING query by an application will
just be passed upstream where it will then be handled by basesrc. Now, if
e.g. totem does a SEEKING query for TIME format and we have a demuxer that
doesn't implement the query, basesrc would answer it with seekable=FALSE in
most cases, and totem can only take that as authoritative answer, not knowing
that the demuxer doesn't implement the SEEKING query. To avoid this, we make
basesrc return FALSE to SEEKING queries in unhandled formats. That way
applications like totem can fall back on assuming seekability depending on
whether a duration is available, or somesuch. Downstream elements doing
such queries are likely to equate an unhandled query with a non-seekable
response as well, so this should be an acceptable fix for the time being.

See #584838, #588944, #589423 and #589424.
2009-07-24 11:54:31 +01:00
Wim Taymans
141e2af580 basesrc: make tag queuing threadsafe
See #588745
2009-07-20 13:40:16 +02:00
Edward Hervey
1d0c6a669f basesrc: Serialize tags into the dataflow. Fixes #588745 2009-07-20 08:49:43 +02:00
Tim-Philipp Müller
79c8e24885 docs: fix API docs for gst_{adapter|byte_reader}_masked_scan_uint32
Clarify byte reader docs a bit: offset is relative to the current
position of the reader, not to the start of the data. Also, the
examples in both the adapter docs and the byte reader docs have
the mask and pattern arguments swapped (see #587561). Spotted
by Carl-Anton Ingmarsson.
2009-07-16 14:17:03 +01:00
Stefan Kost
3ce0c64b45 basetransform: take size once 2009-07-14 08:32:24 +02:00
Stefan Kost
cfae0c15e7 collectpads: make it the best of wims and edwards patch.
Check the right flushing flag, but still add it to the pad-list.
2009-07-07 00:23:41 +01:00
Stefan Kost
2288f9f6ec basetransform: make comment a FIXME comment 2009-07-06 19:52:44 +01:00
Stefan Kost
4bfb312e13 logging: use perf category for dropped buffers 2009-07-06 19:52:44 +01:00
Edward Hervey
e6d949328f collectpads: Don't forward FLUSH_STOP if some input streams are still flushing.
This guarantees that only one FLUSH_STOP event (the last one) will be sent
downstream when a flushing seek is being done through collectpads.
2009-07-05 21:24:42 +02:00
Edward Hervey
ed18ceffe6 collectpads: Update the cookie when setting ourselves as flushing.
This forces the pad status to be re-evaluated on the next _check_pads().
2009-07-04 17:17:18 +02:00
Tim-Philipp Müller
4d76b175ef docs: fix gtk-doc /*< private >*/ marker 2009-07-02 01:18:57 +01:00
Wim Taymans
92b0f32d74 bufferlist: use faster gst_buffer_list_get()
Use the faster gst_buffer_list_get() to get the first buffer of a list.
2009-06-29 11:56:10 +02:00
Ognyan Tonchev
27fe127e76 basesink: take timestamp later
Make sure we don't accidentally cast a bufferlist of a buffer and try to take
the timestamp of it.

Refixes #585960
2009-06-29 11:46:00 +02:00
Stefan Kost
3e21ddf81f logging: add a performace log category
This category can be used to log slow code path and help auditing the
performance. Add FIXME-0.11 to some questionable categories.
2009-06-29 11:28:01 +03:00
Tim-Philipp Müller
31ff42de21 basesink: don't do things with side effects within a g_assert()
Make the bufferlist stuff work properly when things are compiled
with -DG_DISABLE_ASSERT.
2009-06-25 11:25:46 +01:00
Tim-Philipp Müller
6b64d41937 bytereader: add a bunch of utility functions for strings and a data dup function
API: gst_byte_reader_dup_data
API: gst_byte_reader_dup_string
API: gst_byte_reader_dup_string_utf8
API: gst_byte_reader_dup_string_utf16
API: gst_byte_reader_dup_string_utf32
API: gst_byte_reader_skip_string
API: gst_byte_reader_skip_string_utf8
API: gst_byte_reader_skip_string_utf16
API: gst_byte_reader_skip_string_utf32
API: gst_byte_reader_peek_string
API: gst_byte_reader_peek_string_utf8
API: gst_byte_reader_get_string
API: gst_byte_reader_get_string_utf8

And some basic unit tests. Fixes #586568.
2009-06-23 02:17:48 +01:00
Wim Taymans
dd3e2b557e collectpads: use the right flushing flag.
We need to use the pad private flag because the other pad flag is protected with
the pad lock instead.
2009-06-18 16:50:42 +02:00
Edward Hervey
a295d22575 collectpads: Properly handle flushing pads.
If a pad is flushing, it should not be considered as either eos or
containing data.
2009-06-18 16:43:27 +02:00
Wim Taymans
c88e05c861 basesink: add Since tag for new method 2009-06-16 13:34:38 +02:00
Branko Subasic
de5bcfc4cd basesink: add support for buffer list
Fixes #585960
2009-06-16 13:32:37 +02:00
Wim Taymans
6a1d8feccc collectpads: fix .h indentation 2009-06-15 18:44:14 +02:00
Wim Taymans
4c990aab3a basesrc: add some more debug 2009-06-15 18:43:52 +02:00
Tim-Philipp Müller
674447fafe bytereader: add gst_byte_reader_masked_scan_uint32()
Add a pattern scan function similar to the one recently added to
GstAdapter, and a unit test (based on the adapter one).
Fixes #585592.

API: add gst_byte_reader_masked_scan_uint32()
2009-06-13 10:53:49 +01:00
Wim Taymans
25067558dc stepping: more stepping improvements
Update design doc with step-start docs.
Add eos field to step done message
when stepping in reverse, update the segment time field.
Flush out the current step when we are flushing.
2009-06-12 13:20:36 +02:00
Wim Taymans
23b772664b basesink: post step-start
when we clip, also stop the stepping.
Don't do QoS when stepping
Post step-start when queueing and activating the step.
2009-06-12 13:16:29 +02:00
Wim Taymans
6e633cbe33 basesink: update segment start/stop for clipping
When we start stepping, store the start/stop values of the segment before we
install new start/stop values for clipping in non-flushing steps.

for non-flushing steps, update the element start time. For flushing steps, it
does not change because running_time does not advance

Make sure we always perform the stop_stepping operations even when we drop
frames.
2009-06-09 12:06:35 +02:00
Wim Taymans
72e8d9909c basesink: do proper clipping in stepping
Update the stop position of the segment so that we clip correctly.
After clipping in non-flushing mode, rerender the remainder of the buffer.
2009-06-09 10:25:34 +02:00
Wim Taymans
89c42f8506 stepping: do flushing steps correctly
Note in the docs that a flushing step in PLAYING brings the pipeline to the lost
state and skips the data before prerolling again.

Implement the flushing step correctly by invalidating the current step
operation, which would activate the new step operation.
2009-06-08 17:41:56 +02:00
Jan Schmidt
f0f9ed875f basesink: Change awkward wording in a translateable message. 2009-06-08 16:16:27 +01:00
Wim Taymans
a4c6d04b4c basesink: add non-flushing steps
Add support for non-flushing steps and with different rates.
Clear step info when flushing
2009-06-08 16:27:36 +02:00
Mark Nauwelaerts
4bb3702886 basesrc: reply to QUERY_SEEKING with original format. Fixes #584838. 2009-06-05 16:01:27 +02:00
Wim Taymans
fa7db7d247 basesink: fix clipped start/stop after step
Use the segment helpers to get a more accurate clipped start/stop position after
a stepping operation ended.
2009-06-04 17:27:03 +02:00
Wim Taymans
957e40343d basesink: use more correct segment methods
Use the more correct new segment methods for updating the segment before and
after a step.
2009-06-04 12:53:23 +02:00
Wim Taymans
a907f8d145 basesink; handle EOS correctly.
Handle EOS and buffers without a timestamp gracefully.
Remove a warning that is not so much a warning now anymore.
2009-06-03 15:39:13 +02:00
Wim Taymans
c811d51471 basesink: forget previous times when stepping
When we start a flushing step operation, forget about the previous stream time
so that the position reporting works correctly.
2009-06-03 10:39:53 +02:00
Wim Taymans
71a5ebe638 basesink: fix regression in unit tests
Store the timestamp of the buffer after prerolling. While we are prerolled we
want to report the position of the segment start value.
2009-06-02 15:37:22 +02:00
Wim Taymans
484d188440 basesink: post a warning on excessive framedrops
When we go into emergency rendering, post a warning informing the user about
this fact.
2009-06-01 19:37:14 +02:00
Wim Taymans
5479772bec basesink: more stepping in reverse
Fix stepping and position reporting in reverse playback.
2009-06-01 13:37:29 +02:00
Wim Taymans
8c54c70d9c basesink: use start_time as the step start
Use the start_time of the element as the point from where the step operation
starts. This fixes stepping in all paused states.
2009-06-01 13:37:29 +02:00
Wim Taymans
642bb32196 basesink: catch step cases in _wait_preroll()
When a subclass is blocking in _wait_preroll() in the _render method, make sure
we can unlock the subclass and detect this return value from the render method.
2009-06-01 13:37:29 +02:00
Wim Taymans
1caab05238 basesink: more stepping in reverse fixes 2009-06-01 13:37:28 +02:00
Wim Taymans
4be27fa555 basesink: small cleanups 2009-06-01 13:37:28 +02:00
Wim Taymans
0c205b96b4 framestep: implement backwards framestep
Update framestep document, we want to pass the flush flag in the step-done
message.

Add flush flag to the gstmessage.

Update examples to use the new step-done message api.

Implement framestep with playback rates < 0.0 too.
2009-06-01 13:37:28 +02:00
Wim Taymans
1839782379 basesink: add framestepping in time 2009-06-01 13:37:28 +02:00
Wim Taymans
7a7e9629ce basesink: keep track of stepped time
Pass running_time around so that the stepping code can calculate the elapsed
time correctly.
2009-06-01 13:37:27 +02:00
Wim Taymans
f07b86f18b basesink: move stuff around, more stepping
Make start and stop_stepping methods and move their invocation in the right
places.

Perform the atual stepping operation where we have full context about the
timestamps.
2009-06-01 13:37:27 +02:00
Wim Taymans
546c959f26 basesink: first stab at frame stepping in PAUSED
Unlock the prerolled frame and recheck if we need to step.
Keep a simple counter for the frames we're about to skip while stepping and
preroll/post step_done when stepping finished.
2009-06-01 13:37:26 +02:00
Arnout Vandecappelle
c74c3bf1b3 adapter: fix _masked_scan_uint32() at boundaries
gst_adapter_masked_scan_uint32 could return values smaller than offset
if the first byte(s) of the mask are 0 and the pattern matches the
beginning of the adapter.
Added examples to documentation of gst_adapter_masked_scan_uint32().
Also added some more masked boundary tests.
Fixes #584118
2009-05-28 22:02:21 +02:00
Edward Hervey
d72c17e328 gsttypefindhelper: Fix indentation 2009-05-22 12:57:10 +02:00
Wim Taymans
72232cfef1 adapter: improve the flush function
Remove a compare and branch from flush.
2009-05-20 22:20:46 +02:00
Wim Taymans
fd76e4fc93 adapter: potentially save a memcpy in _take
Directly use the assembled_data in _take() functions when we can instead of
copying it out.
2009-05-20 12:48:41 +02:00
Wim Taymans
f3ff7ad055 adapter: micro optimisations 2009-05-20 11:36:11 +02:00
Wim Taymans
49c4e367e9 adapter: avoid comparisions in fast path
Small tweaks to reduce the number of useless compares in loops.
2009-05-20 11:12:43 +02:00
Wim Taymans
9197f4399b adapter: avoid branch in copy code 2009-05-20 10:28:45 +02:00
Wim Taymans
270723c85c adapter: add _masked_scan_uint32
Add a reasonably optimized new gst_adapter_masked_scan_uint32() function
to scan the adapter for a pattern after applying a mask.

Add some unit tests.

API: GstAdapter::gst_adapter_masked_scan_uint32()

Fixes #583187
2009-05-20 00:37:53 +02:00
Wim Taymans
d6b21ba529 adapter: more optimisations
Remove duplicate copy code (_peek_into and _copy) and make a unified
optimized copy function.
2009-05-19 22:13:04 +02:00
Wim Taymans
b23279e922 adapter: don't use realloc, it does a memcpy
Don't use realloc to grow the scratch area because we don't want the memcpy the
old useless data into the new area before we write our new stuff in it.
2009-05-13 23:52:02 +02:00
Wim Taymans
cc7985ab48 adapter: use g_realloc for resizing the buffer
Use g_realloc for resizing the internal buffer instead of a
less fancy _free/_malloc pair.
2009-05-13 22:51:18 +02:00
Wim Taymans
bb013081a4 adapter: move new member to private struct
Move the new members to a private struct because we don't have enough padding
anymore on 32-bits platforms.
2009-05-13 21:35:23 +02:00
Wim Taymans
763a6f6761 adapter: update some docs 2009-05-13 18:50:23 +02:00
Wim Taymans
3d19b75147 adapter: add method to keep track of timestamps
Keep track of the timestamp and offset associated with the current head of the
adapter.

API: GstAdapter::gst_adapter_prev_timestamp()
2009-05-13 16:22:58 +02:00
Wim Taymans
8ceff30ca9 adapter: small cleanups 2009-05-13 16:20:26 +02:00
Wim Taymans
605639a2bf adapter: optimize taking the headbuffer
When a are requested to take a buffer from the adapter that is exactly the
headbuffer, don't make a subbuffer of it but return that head buffer.

Add a unit-test for this new optimisation.
2009-05-12 10:25:40 +02:00
Wim Taymans
50bcf96ffe basesrc: don't ignore pad_start return value 2009-05-11 23:06:39 +02:00
Stefan Kost
e778c5fd8b logging: fix unused variable warning when disabling debug logs.
The var was NULL anyway, bacause of the ifdefs there, the message makes no
sense including it.
2009-04-15 23:35:35 +03:00
Edward Hervey
9e4fa43657 basesink: Remove dead assignments.
sstart/sstop/rstart/rstop are all either:
* assigned values later on before being used in 'do_times:' (EOS and buffers)
* not used (non-EOS events)
2009-04-04 14:39:51 +02:00
Edward Hervey
ec3cd5d438 basesrc: remove dead assignment.
The variable will not be read before it's assigned a value line 942/945
2009-04-04 14:38:52 +02:00
Sebastian Dröge
d12034ab4b gst: Use g_once_init* or G_DEFINE_TYPE 2009-04-04 10:59:39 +02:00
Edward Hervey
471bd93af5 adapter: remove dead assignment.
The value set to to_copy at that line is never used, and is overwritten
further down before being read.
2009-04-03 12:57:24 +02:00
Edward Hervey
6577af3635 basesink : Remove unused variable.
sync is never used anywhere in that code.
2009-04-03 12:57:24 +02:00
Edward Hervey
9bf1d7247a basetransform: move unused variable in the #if 0 block.
That variable is only used by the code which has been if 0'd
2009-04-03 12:57:24 +02:00
Edward Hervey
00cbbc87c7 Remove unused variables detected by LLVM's Clang static analyzer. 2009-04-03 12:56:48 +02:00
Tim-Philipp Müller
13f8041238 basesink: fix once-per-second 'emergency rendering' for case where all buffers but the very first are late
Due to a typo basesink didn't do any emergency rendering of late buffers
if the only buffer ever rendered was the first one with timestamp 0. This
means that in cases where the decoder is very very slow, we'd never see
any buffers but the very first one rendered. Fixes #576381.
2009-03-23 12:34:34 +00:00
Stefan Kost
c2f1cb40fc collectpads: add debug logging to make it easier to trace it 2009-03-15 23:41:33 +02:00
Edward Hervey
c1a75a0b45 gstbasesrc: unsigned long is "%lu", not "%ul". Fixes build on macosx 2009-03-11 11:23:05 +01:00
Stefan Kost
36205e7d42 logging: some additional logging for tracing caps negotiation.
Demote one log that can come quite often. Remove one fixme that is done. Apply
gst-indent changes.
2009-03-10 21:13:40 +02:00
Stefan Kost
de48f9c507 collectpads: revert accidential commit from the queue (me should start using branches) 2009-03-06 22:10:10 +02:00
Stefan Kost
436b0e7275 collectpads: reliably go to eos. Fixes #574160
Update collectpads status when removing pads.
2009-03-06 21:50:19 +02:00
Andy Wingo
e2ff005e63 basesink: propagate UPSTREAM events in pull mode too
* libs/gst/base/gstbasesink.c (gst_base_sink_send_event): Propagate
  upstream events in pull mode too.
2009-03-05 17:42:22 +01:00
Edward Hervey
f0481cb717 GstAdapter: Discard empty buffers in _push(). Fixes #574024 2009-03-04 09:20:43 +01:00
Stefan Kost
df620c9cb6 basesink: move left over handling of the error case to the activate_failed label.
If was left as dead code.
2009-02-26 15:42:06 +02:00
Wim Taymans
79c3c6a339 Error out more specifically on empty caps
When we get empty caps from the getcaps function in the default negotiate
function, post a more descriptive error.
2009-02-23 17:37:46 +01:00
Wim Taymans
805569c873 Don't forward LATENCY event when not ready
When we are not ready to handle a latency query (we are not yet prerolled) we
also don't try to forward the latency event because that might cause unexpected
errors when upstream is not yet linked.
2009-02-23 10:53:17 +01:00
Wim Taymans
d807bca61c GstBaseSink: use new variable to schedule preroll
Use a separate variable to keep track if we need to call the preroll method
instead of abusing the commited variable.
2009-02-03 15:27:34 +01:00
Wim Taymans
fe4233120c GstBaseSink: avoid calling preroll multiple times
Fix a regression introduced by fix for #567725 in commit
1c7ab4ed4f. We should only call the preroll
function once namely when we did not yet commit the state change.

Add a unit test to check that we call the preroll function when interrupting the
clock_wait (see #567725).

Add a unit test to check that we only call the preroll function once.
2009-02-03 12:52:49 +01:00
Thijs Vermeir
8a080229c7 Force reconfigure of basetransform to recheck alloc request
While reconfiguring a basetransform element we need also to recheck
the alloc request. Because it's possible that due to caps changes
the proxy_alloc state is not correct anymore.
(Re-commit after discusion with Wim on IRC)
2009-02-02 18:23:41 +01:00
Wim Taymans
c786c5224a Revert "Check suggested caps for proxy alloc"
This reverts commit 50afd45957.

It breaks the interactive test-scale unit test.
2009-01-30 10:45:17 +01:00
Wim Taymans
a371ea6e5f Revert "Force reconfigure of basetransform to recheck alloc request"
This reverts commit 3a4602d771.

It breaks the interactive test-scale unit test.
2009-01-30 10:45:16 +01:00
Stefan Kost
45a17154b7 Add logging in failure case. Add more details to a todo comment. 2009-01-29 17:46:52 +02:00
Thijs Vermeir
50afd45957 Check suggested caps for proxy alloc
Because we are trying to resolve a suggestion here we don't need
to check on caps for proxy_alloc but we need to check on the suggested
caps instead.
2009-01-29 13:58:55 +01:00
Thijs Vermeir
3a4602d771 Force reconfigure of basetransform to recheck alloc request
While reconfiguring a basetransform element we need also to recheck
the alloc request. Because it's possible that due to caps changes
the proxy_alloc state is not correct anymore.
2009-01-29 13:53:01 +01:00
Thijs Vermeir
3f054f1bae Fix typo 2009-01-25 18:06:44 +01:00
Jan Schmidt
e89ce19497 Update a bunch of gitignores to clean up my git status output 2009-01-23 16:08:40 +00:00
Wim Taymans
8620bc3295 Add new typefing helper function to guess the caps based on the file
extension. See #566661.
API: gst_type_find_helper_for_extension()
2009-01-21 12:48:18 +01:00
Wim Taymans
1c7ab4ed4f Update the last_buffer exactly with the buffer that caused the
preroll and also call the preroll method with that preroll buffer.
Fixes #567725.
2009-01-21 12:43:55 +01:00
Wim Taymans
10f9331125 libs/gst/base/gstbasesink.*: Fix documentation for the wait_clock method, rename basesink -> sink for consistency.
Original commit message from CVS:
* libs/gst/base/gstbasesink.c: (gst_base_sink_commit_state),
(gst_base_sink_wait_clock):
* libs/gst/base/gstbasesink.h:
Fix documentation for the wait_clock method, rename basesink -> sink
for consistency.
2009-01-09 15:43:17 +00:00
Wim Taymans
b2052bb385 libs/gst/base/gstbasesink.c: Release the object lock before calling the query convert pad functions to avoid deadlocks.
Original commit message from CVS:
* libs/gst/base/gstbasesink.c: (gst_base_sink_get_position_last),
(gst_base_sink_get_position_paused), (gst_base_sink_get_position):
Release the object lock before calling the query convert pad functions
to avoid deadlocks.
2009-01-05 15:42:53 +00:00
Edward Hervey
8d18476d94 libs/gst/base/gstcollectpads.c: Use the name of the pads instead of a pointer, helps in debugging.
Original commit message from CVS:
* libs/gst/base/gstcollectpads.c: (gst_collect_pads_available):
Use the name of the pads instead of a pointer, helps in debugging.
2009-01-03 18:39:38 +00:00